Chronic traumatic brain injury in professional soccer players

Worldwide, soccer is the most popular and frequently played sport, with at least 200 million registered participants.¹ Although soccer is considered safe by the general public, the American Academy of Pediatrics has classified soccer as a contact/collision sport.² Accordingly, acute traumatic brain injury (ATBI) is a concern in soccer. It is notable that the concussion rates per 1,000 athlete exposures in American football and men's soccer are equal.³ Furthermore, fatal neurologic injuries in soccer have also been reported. During the period from 1931 to 1974, 26 English soccer players experienced fatal traumatic brain injuries, including eight that were attributed to "heading" the ball.⁴ In the United States from 1979 to 1993, 18 fatal accidents were reported from players running into goalposts.⁵ Another four fatal accidents in high school soccer have also been documented between 1980 and 1988.⁶

In addition to ATBI, chronic traumatic brain injury (CTBI) is of concern in soccer. CTBI, which represents the long-term cumulative neurologic consequence of concussive and subconcussive blows to the head, has been described primarily among professional boxers.^7-10 However, it has been speculated that CTBI can occur in soccer.¹¹ In a survey of British neurologists, a traumatic encephalopathy resembling that encountered in boxing has been reported in soccer players.¹² Tysvaer and Lochen,¹³ Sortland and Tysvaer,¹⁴ and others have observed neuropsychological impairment,¹³ CT pathology,¹⁴ and EEG abnormalities¹⁵ among retired soccer players in Norway. In contrast, a more recent investigation¹⁶ of active soccer players on the United States national team failed to demonstrate any significant neurologic symptomatology (MRI) when compared with control subjects.

There are two distinct mechanisms by which CTBI can occur in soccer. First, because the cognitive impairment from concussion can be cumulative,¹⁷ CTBI can occur from repeated, relatively mild concussions. In soccer, concussions may occur from typical contact sport collisions: head-to-head, head-to-goalpost/ground, and head-to-body (e.g., elbow, foot, knee) collisions. However, unlike some other contact sports(e.g., American football), soccer players do not wear protective headgear.

The second mechanism for potential CTBI is a distinguishing feature of soccer: repetitive heading of the ball. It has been estimated that professional soccer players play approximately 300 division games and head the ball more than 2,000 times during their career.¹⁸ This is compounded by the fact that a plastic-coated soccer ball weighs 396 to 453 g, and the speed of the ball can reach 60 to 120 km/h.¹⁹ It has been calculated that a ball kicked with half power from a distance of 10 m travels 83.2 km/h and hits the head with an impact of 116 kPa. Kicked with full power the ball hits the head with 200 kPa.²⁰ Together these mechanisms suggest that soccer players are particularly vulnerable to head injuries with consequent CTBI.

The current investigation tested the hypothesis that CTBI is associated with professional soccer as evidence by impairment on neuropsychological testing. This study also sought to identify possible risk factors for CTBI(e.g., soccer-related concussions, heading the soccer ball) and to evaluate their contribution to neurocognitive dysfunction in soccer players.

Methods. Subjects. The volunteer group of this investigation consisted of 53 active Dutch male professional soccer team members and 27 male control subjects. The control subjects were members of elite swimming and track teams in The Netherlands. All participants in this study were interviewed and examined using an extensive neuropsychological test battery. Soccer players and control subjects participated in an extensive interview that inquired about age, education, number of general anesthesias, occupational history, number of concussions (sports and nonsports related) in a lifetime, alcohol intake, drug use, and medical conditions that might affect cognitive functioning (table 1). Education was scored on a seven-point scale and, because all participants completed their secondary school education, only the levels 3(technical and vocational training for 12- to 16-year-olds) to 7 (academic level) were used. In addition, the team physicians of the professional soccer players were interviewed concerning the number of concussions (sports and nonsports related) and medical conditions that might interfere with cognitive functioning. All the participants in this study were native Dutchmen who were educated in the Dutch educational system. Nonnative Dutchmen and people who were not educated in the Dutch school system were excluded. Other exclusion criteria included history of drug abuse, epilepsy, and medical conditions that might affect cognitive functioning.

Group comparisons were made between professional soccer players and control subjects for level of cognitive functioning. Intergroup comparisons were made in soccer players who do many "headers" and those who do not(nonheaders). Midfield players and goalkeepers were classified as nonheaders, and forward and defensive players were classified as headers. Trauma index variables regarding the number of concussions with or without loss of consciousness (soccer related), duration of professional soccer career, the number of headers produced each match, and the number of matches in one season were also determined for the professional soccer players. By multiplying the number of headers in each match with the number of matches, we estimated the total number of headers in one season.

Neuropsychological testing. In the current investigation, neuropsychological tests that have proven to be sensitive in the detection of cognitive impairment in contact and collision sports^13,21-23 were used. Each participant in this study received a battery of neuropsychological tests that were administered by a trained psychometrician. The interview was performed by a qualified neuropsychologist. All tests were administered according to standardized instructions and procedures. The test data were scored by a qualified neuropsychologist who was blinded to the status (professional soccer player or control subject) of the participant. The neuropsychological test battery included the following: Raven Progressive Matrices Test(RPM),²⁴ Wisconsin Card Sorting Task,²⁵ Paced Auditory Serial Addition Task (PASAT; slowest version),²⁶ Digit Symbol Test,²⁷ Trail Making A and B,²⁸ Stroop test,²⁹ Bourdon-Wiersma Test,³⁰ subtests of the Wechsler Memory Scale (WMS; Associate Learning, Logical Memory, and Visual Reproduction),³¹ Complex Figure Test(Copy, Immediate Recall, and Delayed Recall),³² 15-Word Learning Test,³³ Benton's Facial Recognition Task,³⁴ Figure Detection Test,³⁵ Verbal Fluency Test,³⁵ and the Puncture Test.³⁶

Statistical analysis. Comparisons of the neuropsychological test scores were performed by calculating the differences between the means using Student's t-test. With a linear regression analysis these differences, their CIs, and one-tailed p values were determined, adjusting for the level of education, alcohol intake, number of general anesthesias, and number of concussions not sustained in soccer. To determine the association between heading-related variables and test scores, unadjusted and adjusted regression coefficients and their p values were calculated using a univariate and a multivariate linear regression model respectively.

Because 27 outcome variables are compared, we performed a Bonferroni correction on the significance level. Furthermore, using the ordinary least squares (OLS) test as described by O'Brien³⁷ and modified by Läuter,³⁸ we tested a global null hypothesis.

Results. Population characteristics. Characteristics of the study population are presented in table 1. The average age of the soccer players was 24.4 ± 4.1 (SD) years compared with 24.5 ± 4.5 years in the control group (p = 0.3). There was no significant difference between the percentage of people sustaining concussions unrelated to soccer (11.3% in the professionals and 14.8% in the control subjects; p = 0.7). None of the professional soccer players in this study abused drugs. Higher alcohol consumption was noted in the control subjects (p = 0.02). Soccer players achieved a higher educational level (p = 0.05) and experienced a greater number of general anesthesias (p = 0.09). Among the 53 professional soccer players, 54% experienced one or more soccer-associated concussions with or without loss of consciousness. During their professional careers 79% of the players sustained one or more head-to-head collisions (with or without postconcussive symptoms).

Professional soccer players reported a median of 800 (range, 50 to 2,100) headers during competitive matches in one soccer season. Headers experienced in training sessions were excluded from this approximation. The median of the number of headers during a match was 16 (range, 1 to 42). Approximately 47% of the players headed the ball 0 to 10 times, 36% headed the ball 11 to 20 times, and 17% headed the ball 21 or more times during an average match. Thirty percent of the players also participated in heading drills during training. These drills consisted of repetitive heading of the ball for 30 minutes once a week.

The median of the number of soccer matches played annually was 50 (range, 25 to 70), and the median of the number of practices per week was six (range, four to nine). Ninety-two percent practiced four to seven times a week and 7.5% practiced more than seven times a week. An average training session lasted 2 hours. Twenty-four percent of the players in this study were forward players, 45% were midfield players, and 30% were defensive players.

Soccer players experienced extensive amateur and professional careers. The median of an amateur soccer career was 12 years (range, 4 to 23 years). Thirty percent of the players played 4 to 10 years, 57% played 10 to 16 years, and 13% played 16 years or longer as amateurs. The median of the duration of a professional soccer career in this study was 5 years (range, 1 to 18 years). Professionally, 43% of the players participated 1 to 5 years, 30% played 6 to 12 years, and 27% played more than 12 years.

Neuropsychological testing. We examined the distributions of the test scores on outliers and skewness. The skew of the distributions necessitated a log transformation for six score-namely, the Bourdon-Wiersma test (number of mistakes), the PASAT, the Facial Recognition Test, the Complex Figure Test (Copy), the Puncture Test (dominant hand), and the total Raven score-to symmetrize and normalize the distribution. The analysis was done with these transformed scores, but in the following the inversely transformed values are presented.

Table 2 shows that professional soccer players exhibited more cognitive impairment when compared with control subjects. Professional soccer players performed poorer on verbal and visual memory, planning, and visuoperceptual processing tasks compared with control subjects. On a number of tests, the differences between the two groups (i.e., the Figure Detection Groninger Intelligentie Test [GIT] subtest, the Complex Figure Test [Copy, Immediate, and Delayed Recall], Logical Memory [WMS], and Visual Reproduction [WMS]) remained significant after adjustments for confounding variables (the number of concussions not related to soccer, alcohol consumption, level of education, and the number of general anesthesias). These psychometric test scores also remained significant after a Bonferroni correction. Furthermore, the one-tailed p value based on the OLS test statistic was 0.00006 (Z_OLS = 3.84). Evidence for a dose-response relation between selected soccer-related exposure variables and the neuropsychological test scores was demonstrated (table 3). An increasing number of headers and concussions incurred during soccer participation were associated negatively with memory, visuoperceptual, and planning capacity. The number of headers was inversely associated with test scores on the Complex Figure Test Immediate (p = 0.048) and Delayed Recall (p = 0.01). The number of concussions sustained in soccer was inversely associated with the test score on the Complex Figure Test Copy(p = 0.01). Field position also influenced performance on neuropsychological testing. Forward and defensive players performed significantly poorer on Figure Detection (p = 0.01), Complex Figure Test Immediate (p < 0.001), Delayed Recall (p < 0.001), Logical Memory (WMS;p = 0.02), and Visual Reproduction (WMS;p = 0.02) compared with midfield players and goalkeepers.

Discussion. These findings suggest that professional soccer may be associated with neurocognitive impairment. The observed impairment in neuropsychological function preferentially involves memory, planning, and visuoperceptual processing. Forward and defensive players tend to be more vulnerable to cognitive impairment because they are more likely to head the ball and experience a higher frequency of soccer-related concussions. Our findings are in agreement with those of Tysvaer and Lochen,¹³ who showed a higher degree of neuropsychological impairment in headers than in nonheaders (20% versus 8%). In their study, 73% of former professional soccer players showed neurocognitive impairment (3% severe, 38% moderate, and 32% mild). The neuropsychological differences between the retired players and the control subjects were documented in tasks regarding perception, learning, and mental speed. The players in the study by Tysvaer and Lochen¹³ had an average age of 48.6 years (range, 35 to 64 years), and were retired.

In a neurologic evaluation of 20 soccer players on the United States national team,¹⁶ no association was found between heading the soccer ball and neurologic symptoms, including MRI-detected abnormalities. The discrepancy between this investigation and our findings may be related to methodologic issues. This study did not use the more sensitive neuropsychological techniques nor did it distinguish between different styles of soccer play in training and competition.

The Complex Figure Test proved to be sensitive in detecting symptoms of neurocognitive impairment in professional soccer players. The sensitivity of the Complex Figure Test may be explained by the fact that this test requires memory, planning, and visuospatial ability (i.e., cognitive functions that are sensitive to deterioration after concussion and multiple trivial impacts). According to the standard of Osterieth,³² as used in clinical practice, 7% of the control subjects and 45% of the professional soccer players showed moderate to severe impaired scores. Surprisingly, the PASAT did not differentiate between soccer players and control subjects. In this study we utilized the slowest PASAT version, which could explain the lack of sensitivity.

In our cross-sectional study, several methodologic precautions were taken to minimize the problems associated with group comparisons. All the participants in this study were native Dutchmen who were educated according to the Dutch educational system, therefore eliminating misinterpretations caused by educational differences. In addition, drug users and participants with medical conditions that might interfere with cognitive functioning were excluded from this investigation.

Other potential biases associated with physical exertion and exercise in soccer were excluded by using elite swimmers and runners as control subjects. To ensure reliability, tests and interviews were administered by a certified neuropsychologist and trained psychometricians. The test scores were scored by a neuropsychologist who was not informed about the status (professional soccer player or control subject) of the participants. Confounding variables such as alcohol use, the number of concussions not sustained in soccer, the level of education, and the number of general anesthesias were adjusted using multivariate regression analysis. The problem with multiple end point testing was controlled by a Bonferroni correction of the significance level and a global null hypothesis testing procedure.

Difference in mentality and test attitude could also explain these results. However, soccer players had a higher level of education, and all participants were tested by a trained psychometrician. Nevertheless, there may be inherent differences between those who choose to play soccer and those who choose to swim or run that may affect neuropsychological test performance and that may not be captured adequately in the variables we controlled. Our finding that professional soccer is associated with neuropsychological impairment supports the hypothesis that CTBI is a health problem for soccer players. Although the neuropsychological impairment is milder than what might be anticipated in boxing, the dose-response relation between traumatic brain injury (i.e., concussions and heading the ball) and cognitive impairment presents a medical and public health concern. Accordingly, surveillance and prevention measures should be undertaken to maximize safety. For example, soccer players should have a medical evaluation after sustaining a concussion. Baseline neuropsychological testing (performed in preseason) and postconcussion sideline follow-up evaluations may be useful in determining the effects of concussion and the appropriateness for return to competition. To limit the adverse effects of heading the ball, precautionary measures also need to be established. In other contact sports with similar concussion rates(e.g., American football³), these preventive measures are recommended.

Participation in professional soccer may affect adversely some aspects of cognitive functioning (i.e., memory, planning, and visuoperceptual processing). This impairment in neurocognitive function appears to be attributed to increased heading of the ball and soccer-related concussions experienced by forward and defensive players. Whether our findings can be extrapolated to amateur or lower exposure soccer players remains to be determined in future investigations.

Acknowledgment

We thank H.J.A. Schouten, PhD, for his helpful statistical advice.